Page 3 of 3

Re: Applying AI to classify the ships of Oolite

Posted: Mon Mar 14, 2011 12:02 pm
by Disembodied
TGHC wrote:
And we've already got variable fuel costs for different types of ships:
Oooooh I didn't know that, how is that working then?
It's in this thread here ...

Re: Applying AI to classify the ships of Oolite

Posted: Mon Mar 14, 2011 3:20 pm
by Makandal
Disembodied wrote:
Commander McLane wrote:
My whole point is that no such thing as a 'correct price' exists. Contra-factually pretending that it does can only lead to confusion and faux-debates.
But if the whole thing is contained within a game context, e.g. within the pages of What Starship? or a report by Makandal's Intergalactic Mammaloid Consumers' Association, then any resulting in-game confusion and debate are entirely proper, and add to the game-in-the-head. A Fer-de-Lance is hugely over-priced, by the statistics; what you're paying for is the megawalnut veneer and the aura of exclusivity. I mean, compare the reverberating clang when a Python's airlock bangs shut with the whisper-soft thunk you get when you tighten up your Ferdy: it's brute-force mechanical compared to hand-crafted perfection, and that's what you're paying for.
Makandal would like to state he is NOT a member of IMCA and will never support, sign, preface any report from the IMCA.
With the reaction I see about the pricing, I think Makandal will, with utmost caution, move out of this touchy and sensitive subject.

Re: Applying AI to classify the ships of Oolite

Posted: Mon Mar 14, 2011 3:50 pm
by Disembodied
Makandal wrote:
Makandal would like to state he is NOT a member of IMCA and will never support, sign, preface any report from the IMCA.
With the reaction I see about the pricing, I think Makandal will, with utmost caution, move out of this touchy and sensitive subject.
I think it's just a matter of presentation. ;) I don't see anything wrong with at least giving designers some sort of idea about what sort of price band a ship might be suited for. Traditionally there should be eight – something like: Economy, Value, Inexpensive, Mid-Range, Expensive, High, Exclusive and Extreme.

Re: Applying AI to classify the ships of Oolite

Posted: Mon Mar 14, 2011 4:13 pm
by Zireael
Disembodied wrote:
Makandal wrote:
Makandal would like to state he is NOT a member of IMCA and will never support, sign, preface any report from the IMCA.
With the reaction I see about the pricing, I think Makandal will, with utmost caution, move out of this touchy and sensitive subject.
I think it's just a matter of presentation. ;) I don't see anything wrong with at least giving designers some sort of idea about what sort of price band a ship might be suited for. Traditionally there should be eight – something like: Economy, Value, Inexpensive, Mid-Range, Expensive, High, Exclusive and Extreme.
Seconded.

Re: Applying AI to classify the ships of Oolite

Posted: Mon Mar 14, 2011 5:14 pm
by Makandal
That was the intention. I might be a bit rude sometimes in my presentation. I am a geologist, not a psychologist.

I forgot to present 2 algorithms called "Association Rules Algorithm" and CN2. They are pretty interesting because it uses "natural language". Some rules are pretty straight forward:
If Cargo bay extension<=1.00 AND Military shields available=Yes, Shield boosters available=Yes

Other are more subtle:
If EnergyRechargeRate>3 AND MilitaryShieldAvailable=1, MaxSpeed[0.359, 0.373]
If EnergyRechargeRate>3 AND MissileSlots>3, Cargo_Capacity[17, 45]


Once more, this is not a firm guideline to build ships. I advise you to read this article on allometry laws of flight, the approach is similar.
These are just pretty hard trends emerging from the 160 ships I have on my list (All these rules are supported by more than 25% of the ships.).

Re: Applying AI to classify the ships of Oolite

Posted: Tue Mar 15, 2011 8:10 am
by TGHC
It's always interesting and sometimes exciting to see a completely new idea. With the breadth and depth of creativity on this board someone will no doubt be able to make some good use of this analysis.